供应链管理一二

单击此处编辑母版文本样式,2,第三级,第四级,第五级,*,Chapter 2,Inventory Management, Supply Contracts and Risk Pooling,Zhang Yuhua,Professor of Guangdong University of Foreign Studies,Tel: 020-39328850,Outline of the Presentation,2,.1 Introduction to Inventory Management,2,.2 Single Warehouse Inventory,(1),EOQ,(2) Demand Forecast,(3) Supply Contracts,(4),A multi-Period Inventory Model,(5) Periodic Review Policy,2,.3 Risk Pooling,2,.4 Centralized vs. Decentralized Systems,2,.5,Managing Inventory in the SC,2,.6 Practical Issues in Inventory Management,2.1 Inventory,Where do we hold inventory?,Suppliers and manufacturers,warehouses and distribution centers,retailers,Types of Inventory,WIP,raw materials,finished goods,Why do we hold inventory?,Unexpected changes in customer demand,The presence in many situations of a significant uncertainty,Economies of scale offered by transportation companies,Lead Time, Capacity limitations,Understanding Inventory,The inventory policy is affected by:,Demand Characteristics,Lead Time,Number of Products,Objectives,Service level,Minimize costs,Cost Structure,Cost Structure,Order costs,Fixed,Variable,Holding Costs,Insurance,Maintenance and Handling,Taxes,Opportunity Costs,Obsolescence,2.2.1 EOQ: A Simple Model*,A Case :,Book Store Mug Sales,Demand is constant, at 20 units a week,(D for a year),Fixed order cost of $12.00, no lead time,(k),Holding cost of 25% of inventory value annually,(H),Mugs cost $1.00, sell for $5.00,Question,How many,(Q), when to order?,EOQ illustrates the trade-offs between ordering and storage costs.,EOQ: A View of Inventory*,Time,Inventory,Order,Size,Note:, No Stockouts, Order when no inventory, Order Size determines policy,Avg. Inven,EOQ: Calculating Total Cost*,Purchase Cost Constant,Holding Cost:,(Avg. Inven) * (Holding Cost),Ordering (Setup Cost):,Number of Orders * Order Cost,Goal:,Find the,Order Quantity,that Minimizes These Costs,EOQ:Total Cost*,Total Cost,Order Cost,Holding Cost,EOQ: Optimal Order Quantity*,Optimal Quantity = (2*Demand*Setup Cost)/holding cost,Q=Sqrt(2*D*K)/H)=Sqrt(2*20*52*12)/25%)=316,So for our problem, the optimal quantity is 316,EOQ,总成本,用,用如下,公,公式表,示,示：,等式右,边,边的第,一,一项表,示,示的是,库,库存持,有,有成本,。,。右边,的,的第二,项,项表示,的,的是订,货,货成本,或,或者生,产,产准备,成,成本，,R/Q,表示每,年,年向供,货,货点发,出,出订货,定,定单的,次,次数。,由,由此可,见,见，如,果,果,Q,增加，,而,而每年,的,的需求,固,固定，,那,那么每,年,年的订,货,货的次,数,数就相,应,应减少,。,。第三,项,项是货,物,物自身,的,的成本,，,，它不,影,影响最,优,优订货,批,批量和,最,最优订,购,购周期,的,的确定,。,。,为获得,使,使总成,本,本最低,的,的最优,订,订货批,量,量,Q*,，总成,本,本,TC,看作以,Q,为自变,量,量的函,数,数，将,TC,函数对,Q,微分：,EOQ,令为,0,，,可求出,最,最优的,Q,*,为,每年的,订,订货次,数,数为,两次订,货,货之间,最,最佳的,时,时间间,隔,隔为,EOQ: ImportantObservations,*,Tradeoffbetween set-upcostsandholdingcostswhen determining order quantity.,TotalCost is,notparticularly sensitive,totheoptimalorderquantity,TheEffect ofDemandUncertainty,Most companiestreattheworldasifitwerepredictable:,Production andinventoryplanningare based on forecastsofdemandmadefarinadvanceofthesellingseason,Companiesare aware of demanduncertaintywhen theycreate aforecast, butthey designtheirplanningprocess as if theforecasttrulyrepresents reality,Recenttechnologicaladvanceshave increasedthelevelofdemand uncertainty:,Shortproduct lifecycles,Increasing productvariety,DemandForecast,Thethreeprinciplesofallforecastingtechniques:,Forecastingisalwayswrong,The longertheforecast horizon, theworst theforecast,Aggregateforecastsare more accurate,Case:SnowTimeSporting Goods,Fashion items have shortlifecycles, high varietyof competitors,SnowTime Sporting Goods,New designs arecompleted,One productionopportunity,Basedon past sales,knowledgeof the industry, and economicconditions, the marketing department has a probabilistic forecast,The forecast averages about 13,000,butthereis achance that demandwillbe greateror less than this.,Case2,Supply Chain Time Lines,Jan 00,Jan 01,Jan 02,Feb 00,Sep00,Sep01,Design,Production,Retailing,Feb01,Production,SnowTime DemandScenarios,SnowTime Costs,The variable Production cost per unit (C):$80,Selling price per unit (S):$125,Salvage value per unit (V):$20,Fixedproduction cost (F):$100,000,To start production,themanufacturer has toinvest $100,000independent oftheamount produced.,Q isproductionquantity,D demand,Profit =Revenue -Variable Cost -Fixed Cost + Salvage,SnowTime Scenarios,Scenario One:,Suppose you make 12,000 jackets anddemand ends upbeing 13,000 jackets.,Profit =125(12,000) - 80(12,000) - 100,000 =$440,000,Scenario Two:,Suppose you make 12,000 jackets anddemand ends upbeing 11,000 jackets.,Profit = 125(11,000) - 80(12,000) - 100,000 + 20(1000) =$ 335,000,Profit =Revenue -Variable Cost -Fixed Cost + Salvage,SnowTime Best Solution,Findorderquantitythatmaximizesweighted average profit.,Question:Willthisquantity be less than, equal to, orgreater than average demand?,Whatto Make?,Question:Willthisquantitybe less than, equalto, or greaterthanaverage demand?,Average demandis 13,100,Lookat marginal cost Vs.marginalprofit,if extra jacketsold, profit is 125-80 =45,if not sold, cost is80-20 = 60,So wewillmakelessthanaverage,SnowTime Expected Profit,The quantity that maximizes averageprofit, is about 12,000.,Whatto Make?,If the cost ofnot selling anadditionalunitis largerthantheprofit from sellingan additional unit,the optimal quantityin general will belessthanaverage demand,while ifthe reverse istrue,theoptimal order quantity ingeneral will begreater thanaverage demand.,SnowTime Expected Profit,It indicates that producing 9,000 unitsor producing 16,000unitswillleadto aboutthe same average profit of $294,000.,SnowTime:,ImportantObservations,Tradeoffbetween ordering enough to meet demand and ordering toomuch,Several quantities have the same averageprofit,Average profitdoesnot tell the whole story,Question:9000and 16000unitslead toaboutthesameaverage profit, sowhichdo we prefer?,SnowTimeExpected Profit,Probability ofOutcomes,Whenthe productionquantity is 16,000 units,thedistribution ofprofit isnotsymmetrical. Lossesof $220,000 happen about11%,whileprofits of atleast$410,000happen 50%.,Whenthe productionquantity is 9,000 units ,the distribution hasonlytwopossible outcomes. Profitis either$200,000withprobability ofabout11%,or $305,000 with probability of about 89%.,Key Insights from this Model,The optimal order quantity isnot necessarilyequal toaverage, or forecastdemand,The optimal quantitydepends on therelationship between marginalprofit and marginalcost,As order quantity increases, average profit firstincreasesand then decreases,As production quantity increases, risk increases.In otherwords, theprobability oflarge gains and oflargelosses increases,SnowTimeCosts: TheEffect ofInitial Inventory,Productioncostperunit(C):$80,Selling price per unit (S): $125,Salvage value per unit (V): $20,Fixedproduction cost (F): $100,000,Qisproductionquantity,Ddemand,Profit=,Revenue-VariableCost-FixedCost+Salvage,SnowTimeExpectedProfit,InitialInventory,Supposethatoneofthejacketdesignsisamodelproducedlastyear.,Someinventoryisleftf,Assume the same demand pattern as before,If only old inventory is sold, no setup cost,Question:,If there are 5000 units remaining, what should SnowTime do? What should they do if there are 10,000 remaining?,InitialInventoryandProfit,InitialInventoryandProfit,InitialInventoryandProfit,Ifthemanufacturerdoesnotproduceanyadditionalsuits,nomorethan5,000unitscanbesoldandnoadditionalfixedcostwillbeincurred.However,itthemanufacturerdecidestoproduce,afixedproductioncostischargedindependentoftheamountproduced.,InitialInventoryandProfit,Averageprofitexcludingfixedproductioncost,Averageprofitincludingfixedproductioncost,(1)thereare5000unitsremaining,Ifnothingisproduced,averageprofitisequalto625000.,Productionshouldincreaseinventoryfrom5,000unitsto12,000units.Thus,averageprofitisequalto771000(fromthefigure).,(2) thereare 10,000units remaining,It iseasyto see that there is noneedto produceanythingbecause the averageprofit associated with aninitial inventory of 10,000 is larger than what we would achieveif we produceto increase inventory to12,000 units.,If weproduce,the most we canmakeon average isa profit of$375,000.Thisis the same averageprofit that wewillhaveif our initialinventoryis about8,500units.,Hence,if our initialinventoryis below 8,5000units, weproduce to raise the inventory level to12,000 units.If initialinventoryis at least 8,5000units,we shouldnotproduce anything.,Analysis,(s, S) Policies,For some starting inventory levels,it is better to notstart production,If westart, wealways produceto the same level,Thus,we use an(s,S) policy.If the inventory level is below s,we produce upto S.,sis thereorder point, andSis theorder-up-to level,The differencebetween the twolevels isdriven bythefixedcosts associated with ordering, transportation, or manufacturing,Manufacturer,Manufacturer DC,Retail DC,Stores,Fixed Production Cost,=$100,000,Variable Production Cost,=$35,Selling Price=$125,Salvage Value=$20,WholesalePrice=$80,2.2.3Supply Contracts,Who takesthe risk?Whatwould themanufacturer like?,Demand Scenarios,DistributorExpected Profit,DistributorExpected Profit,470,000,Supply Contracts (cont.),Distributoroptimal order quantityis 12,000units,Distributor expectedprofit is$470,000,Manufacturer profitis $440,000,Supply Chain Profitis $910,000,Is there anything that the distributor and manufacturercando to increasetheprofit ofboth?,Supply,Contracts,Buy-Back Contracts,The selleragrees tobuybackunsold goods from the buyer for some agreed-upon pricehigher than thesalvage value,Manufacturer,Manufacturer DC,Retail DC,Stores,Fixed Production Cost =$100,000,Variable Production Cost=$35,Selling Price=$125,Salvage Value=$20,WholesalePrice=$80,Supply Contracts,(between manufacturer andretailer),In the previousstrategy,theretailer takesall the risk and themanufacturer takeszerorisk.Thisis why the retailerhasto beveryconservative with the amounthe orders.,If the retailercantransfersome oftherisk tothemanufacturer, the retailermaybe willing to increase his orderquantity and thus increasebothhisprofitandthemanufacturerprofit.,RetailerProfit(Buy Back=$55),RetailerProfit(Buy Back=$55),$513,800,*,假定制造,商,商同意以,55,美元的价,格,格从零售,商,商处购买,销,销售不了,的,的衣服，,则,则零售商,愿,愿意增加,订,订货量到,14000,件，并获,得,得,513800,美元的利,润,润，而制,造,造商的平,均,均利润增,加,加到,471900,ManufacturerProfit(Buy Back=$55),ManufacturerProfit(BuyBack=$55),$471,900,Supply,Contracts,Revenue-Sharing,Contracts,Thebuyer sharessome ofits revenue with the seller, in returnfora discounton the wholesaleprice,Manufacturer,Manufacturer DC,Retail DC,Stores,Fixed Production Cost =$100,000,Variable Production Cost=$35,Selling Price=$125,Salvage Value=$20,Wholesale Price=$?,Supply Contracts,(betweenmanufacturerswholesale priceandretailer),Whatdoes wholesaleprice drive?,Howcanmanufacturerbenefitfrom lowerprice?,RetailerProfit,(Wholesale Price$70, Retailersback15%tomanufacturer),RetailerProfit(WholesalePrice $70, RS 15%),$504,325,ManufacturerProfit(WholesalePrice $70, RS 15%),ManufacturerProfit(WholesalePrice $70, RS 15%),$481,375,Supply Contracts,Wholesale Price$70,RS15%,Supply,Contracts,Global Optimization,Contracts,Thisunbiased decision makerwould considerthetwosupply chainpartners, the supplierandthebuyer, as two members ofthesame organization.Thatis,thetransfer ofmoney between the parties is ignored and the unbiased decisionmaker will maximizesupply chainprofit,Manufacturer,Manufacturer DC,Retail DC,Stores,Fixed Production Cost =$100,000,Variable Production Cost=$35,Selling Price=$125,Salvage Value=$20,Wholesale Price=$80,Supply Contracts,Whatisthemaximumprofit thatthesupply chaincanachieve? Toanswerthisquestion, one needstoforget aboutthetransfer ofmoney fromtheretailertothemanufacturer.,Supply ChainProfit,Supply ChainProfit,(Globaloptimization),$1,014,500,Supply Contracts,Supply Contracts: Key Insights,Effective supplycontracts allowsupplychain partners to replace,sequential optimization,by,global optimization,BuyBackandRevenueSharingcontracts achieve thisobjective through,risksharing,Supply Contracts: Case Study,Example:Demandfora movienewly released videocassette typicallystarts highanddecreases rapidly,Peakdemandlastabout 10 weeks,Blockbusterpurchases acopyfrom astudio for $65 and rentfor$3,Hence,retailermustrentthetapeatleast22timesbeforeearningprofit,Retailerscannotjustifypurchasingenoughtocoverthepeakdemand,In1998,20%ofsurveyedcustomersreportedthattheycouldnotrentthemovietheywanted,SupplyContracts:CaseStudy,Startingin1998Blockbusterenteredarevenuesharingagreementwiththemajorstudios,Studiocharges$8percopy,Blockbusterpays30-45%ofitsrentalincome,EvenifBlockbusterkeepsonlyhalfoftherentalincome,thebreakevenpointis6rentalpercopy,TheimpactofrevenuesharingonBlockbusterwasdramatic,Rentalsincreasedby75%intestmarkets,Marketshareincreasedfrom25%to31%(The2ndlargestretailer,HollywoodEntertainmentCorphas5%marketshare),OtherContracts,QuantityFlexibilityContracts,Supplierprovidesfullrefundforreturneditemsaslongasthenumberofreturnsisnolargerthanacertainquantity,SalesRebateContracts,Supplierprovidesadirectincentivetotheretailertoincreasesalesbymeansofarebatepaidbythesupplierforanyitemsoldaboveacertainquantity,Contractsformake-to-stock/make-to-ordersupplychains,Pay-backContracts,Inthiscontract,thebuyeragreestopaysomeagreed-uponpriceforanyunitproducedbythemanufacturerbutnotpurchasedbythedistributor.,Cost-sharingContracts,In acost-sharing contracts, the buyer sharessomeof the production costwiththe manufacturer, inreturn for a discount onthewholesaleprice.,Contractswithasymmetricinformation,Capacity reservationcontracts,Manufacturer pays toreserve acertain levelof capacity with thesupplier.,Advance purchase contracts,Supplier charges theadvancepurchasepriceforfirms orders placedprior tobuilding capacity and a different pricefor any additional orderplaced when demand is realized.,Contractsfor nonstrategic components,Long-termContracts,Alsocalled forwardor fixed commitmentcontracts,these contracts specifya fixed amountof supplyto be delivered atsomepointin the future.Thesupplier and the buyer agreeson both the price and quantity tobe delivered tothebuyer.,Flexible,or option,Contracts,The buyerprepays arelativelysmall fractionof the productprice upfront,in returnfora commitment from the supplierto reserve capacityup to a certain level.,Contractsfor nonstrategic components,Spotpurchase,Buyers look foradditional supply in theopenmarket.,PortfolioContract,To find aneffectivecontracts, thebuyer needs toidentifythe appropriatemixof low-price yet low-flexibility (Long-term)contracts,reasonable price but better flexibility(option)contracts,or unknown price and quantitysupply but nocommitment(Spot market).,2.2.4A Multi-PeriodInventoryModel,Often, there are multiplereorder opportunities,Consider acentral distribution facilitywhich orders from amanufacturer and delivers toretailers.Thedistributor periodicallyplaces ordersto replenish its inventory,The DC holds inventory to:,Satisfy demandduring lead time,Protect againstdemand uncertainty,Balance fixed costsand holding costs,Reminder:,The NormalDistribution,Average =30,Standard Deviation =5,Standard Deviation =10,The Multi-Period Continuous ReviewInventoryModel,Normally distributedrandom demand,Fixedorder cost plus a cost proportional toamount ordered.,Inventorycostis chargedperitemper unit time,If anorder arrivesand thereis noinventory, theorder islost,The distributorhasa requiredservice level.This isexpressedas the thelikelihood that thedistributor will not stock out during lead time.,Intuitively,howwillthiseffectourpolicy?,AViewof(s,S)Policy,Time,InventoryLevel,S,s,0,Lead,Time,Lead,Time,InventoryPosition,The(s,S)Policy,(s,S)Policy:Whenevertheinventorypositiondropsbelowacertainlevel,s,weordertoraisetheinventorypositiontolevelS.,Thereorderpointisafunctionof:,TheLeadTime,Averagedemand,Demandvariability,Servicelevel,Notation,AVG=averagedailydemand,STD=standarddeviationofdailydemand,LT=replenishmentleadtimeindays,h=holdingcostofoneunitforoneday,K=fixedcost,SL=servicelevel(forexample,95%).Thisimpliesthattheprobabilityofstockingoutis100%-SL(forexample,5%),Also,theInventoryPositionatanytimeistheactualinventoryplusitemsalreadyordered,butnotyetdelivered.,Analysis,Thereorderpoint(s)hastwocomponents:,Toaccountforaveragedemandduringleadtime:LT,AVG,Toaccountfordeviationsfromaverage(wecallthissafetystock),z,STD,LT,where,z,ischosenfromstatisticaltablestoensurethattheprobabilityofstockoutsduringleadtimeis,100%-SL.,Sincethereisa fixed cost, we order morethanuptothereorderpoint:,Q=,(2,K,AVG)/h,Thetotalorder-up-tolevelis:S=Q+s,Example,Thedistributorhas historicallyobservedweeklydemand of:AVG=44.6STD= 32.1Replenishment leadtimeis2 weeks,anddesiredservice levelSL= 97%,Average demandduringleadtime is:44.6, 2=89.2,SafetyStockis:,1.88, 32.1,2=85.3,Reorder point is thus175,orabout3.9weeksofsupply at warehouseandinthe pipeline,What is Reorderpoint?,what is theorder-up-to-level?,Example,Cont.,Weeklyinventory holdingcost:,.87,Therefore, Q=679,Order-up-tolevelthus equals:,Reorder Point +Q= 176+679=855,Suppose thedistributorplacesorders every month,What policyshouldthe distributor use?,What about thefixedcost?,Base-Stock Policy,Inventory Level,Time,Base-stock Level,0,Inventory,Position,r,r,L,L,L,PeriodicReviewPolicy,Each reviewechelon,inventorypositionisraised to thebase-stock level.,Thebase-stocklevelincludestwocomponents:,Average demandduring,r+L,days (thetimeuntilthenextorderarrives):,(r+L)*AVG,Safetystockduring thattime:,z,*STD*,r+L,2.3,Risk Pooling,Coefficientofvariation,Coefficientofvariation= Standard deviation/Average demand,The,Standarddeviationmeasuresthe absolute variability of customer demands, the,coefficientofvariation,measuresvariabilityrelativetoaverage demand,.,2.3RiskPooling Example,Considerthesetwosystems:,Some problems faced by ACME, acompanythatproducesanddistributeselectronicequipmentinthe NortheastoftheUnited States.,(1)Thecurrentdistributionsystem partitionsthe Northeastinto twomarkets,each of which hasa singlewarehouse.Retailersreceiveitemsdirectlyfrom thewarehouses;each retailer is assigned to asingle marketandreceivesdeliveriesfrom thecorresponding warehouse.,Supplier,Warehouse One,Warehouse Two,Market One,Market Two,Risk PoolingExample(cont),(2)Replace thetwowarehouseswith asingle warehouse.The sameservicelevel,97%, be maintainedregardlessofthelogistics strategy employed.,This systemallowsACMEtoachieve eitherthesameservice level of 97%with muchlowerinventory or ahigher servicelevelwiththesameamountoftotalinventory.,Market Two,Supplier,Warehouse,Market One,Risk PoolingExample(cont),Forthe sameservicelevel, which systemwill requiremoreinventory?Why?,Forthe sametotalinventory level, which systemwill havebetter service?Why?,What arethefactorsthat affecttheseanswers?,Risk PoolingExample(cont),Compare thetwosystems:,twoproducts(A,B),maintain97%servicelevel,$60ordercost,$.27 weeklyholding cost,$1.05transportationcost perunit in decentralizedsystem,$1.10incentralizedsystem,1 weekleadtime,Table1Historical Datafor ProductAandB,Risk PoolingExample,Thetables includeweekly demandinformationforeachproduct forthelasteightweeksineach marketarea.Observe thatProductB is aslow-movingproduct,thedemandfor ProductBisfairlysmallrelativetothe demandforProductA.,Risk PoolingExample,Table 2Summaryof Historical Data,Product,Average Demand,Standard Deviation Demand,Coefficient of Variation,Market1,A,39.3,13.2,0.34,Market1,B,1.125,1.36,1.21,Market2,A,38.6,12.0,0.31,Market2,B,1.25,1.58,1.26,Total,A,77.9,20.71,0.27,Total,B,2.375,1.9,0.81,RiskPoolingExample,Table 3Inventory Levels,Product,AVG D,Safety Stock,Reorder point,Q,Order-up to level,Market1,A,39.3,25.8,65,132,197,Market1,B,1.125,2.58,4,25,29,Market2,A,38.6,22.8,62,131,193,Market2,B,1.25,3,5,24,29,Total,A,77.9,39.35,118,186,304,Total,B,2.375,3.61,6,33,39,RiskPooling:ImportantObservations,Centralizinginventory control reduces bothsafety stockandaverageinventory levelforthesame service level.,Thehigher the coefficient of variation,thegreaterthebenefitobt